如何将函数作为参数传递给 OpenSCAD 模块?
How do I pass a function as a parameter to an OpenSCAD module?
在过去的几天里,我对使用基于编程语言的软件创建 3D 模型的想法产生了兴趣。我一直在玩的一种语言是 OpenSCAD,事实证明它对创建有趣的形状非常有帮助。
我目前正在尝试使用 OpenSCAD 创建一朵花,我遇到了一个问题,我无法使用我在网上找到的文档或其他资源来规避。
这里是问题的简短形式:
我可以将函数作为参数传递给 OpenSCAD 模块吗?
如果可以,怎么做?如果没有,为什么不呢?我可以做些什么呢?
这让我想到了问题的长篇形式,并根据我的情况做了具体说明:
我正在尝试使用 2D 极坐标函数的线性拉伸创建花瓣,并将其与 3D 函数相交。
为此,我从 http://spolearninglab.com/curriculum/lessonPlans/hacking/resources/software/3d/openscad/openscad_math.html 上找到的两个非常好的模块开始。我并没有声称自己一开始就写了它们。
首先 - Dan Newman 的 3D 绘图仪
/* 3Dplot.scad */
// 3dplot -- the 3d surface generator
// x_range -- 2-tuple [x_min, x_max], the minimum and maximum x values
// y_range -- 2-tuple [y_min, y_max], the minimum and maximum y values
// grid -- 2-tuple [grid_x, grid_y] indicating the number of grid cells along the x and y axes
// z_min -- Minimum expected z-value; used to bound the underside of the surface
// dims -- 2-tuple [x_length, y_length], the physical dimensions in millimeters
//Want to pass in z(x,y) as parameter
module 3dplot(x_range=[-10, +10], y_range=[-10,10], grid=[50,50], z_min=-5, dims=[80,80]){
dx = ( x_range[1] - x_range[0] ) / grid[0];
dy = ( y_range[1] - y_range[0] ) / grid[1];
// The translation moves the object so that its center is at (x,y)=(0,0)
// and the underside rests on the plane z=0
scale([dims[0]/(max(x_range[1],x_range[0])-min(x_range[0],x_range[1])),
dims[1]/(max(y_range[1],y_range[0])-min(y_range[0],y_range[1])),1])
translate([-(x_range[0]+x_range[1])/2, -(y_range[0]+y_range[1])/2, -z_min])
union()
{
for ( x = [x_range[0] : dx : x_range[1]] )
{
for ( y = [y_range[0] : dy : y_range[1]] )
{
polyhedron(points=[[x,y,z_min], [x+dx,y,z_min], [x,y,z(x,y)], [x+dx,y,z(x+dx,y)],
[x+dx,y+dy,z_min], [x+dx,y+dy,z(x+dx,y+dy)]],
faces=prism_faces_1);
polyhedron(points=[[x,y,z_min], [x,y,z(x,y)], [x,y+dy,z_min], [x+dx,y+dy,z_min],
[x,y+dy,z(x,y+dy)], [x+dx,y+dy,z(x+dx,y+dy)]],
faces=prism_faces_2);
}
}
}
}
第二个 - 2D Grapher
/* 2dgraphing.scad */
// function to convert degrees to radians
function d2r(theta) = theta*360/(2*pi);
// These functions are here to help get the slope of each segment, and use that to find points for a correctly oriented polygon
function diffx(x1, y1, x2, y2, th) = cos(atan((y2-y1)/(x2-x1)) + 90)*(th/2);
function diffy(x1, y1, x2, y2, th) = sin(atan((y2-y1)/(x2-x1)) + 90)*(th/2);
function point1(x1, y1, x2, y2, th) = [x1-diffx(x1, y1, x2, y2, th), y1-diffy(x1, y1, x2, y2, th)];
function point2(x1, y1, x2, y2, th) = [x2-diffx(x1, y1, x2, y2, th), y2-diffy(x1, y1, x2, y2, th)];
function point3(x1, y1, x2, y2, th) = [x2+diffx(x1, y1, x2, y2, th), y2+diffy(x1, y1, x2, y2, th)];
function point4(x1, y1, x2, y2, th) = [x1+diffx(x1, y1, x2, y2, th), y1+diffy(x1, y1, x2, y2, th)];
function polarX(theta) = cos(theta)*r(theta);
function polarY(theta) = sin(theta)*r(theta);
module nextPolygon(x1, y1, x2, y2, x3, y3, th) {
if((x2 > x1 && x2-diffx(x2, y2, x3, y3, th) < x2-diffx(x1, y1, x2, y2, th) || (x2 <= x1 && x2-diffx(x2, y2, x3, y3, th) > x2-diffx(x1, y1, x2, y2, th)))) {
polygon(
points = [
point1(x1, y1, x2, y2, th),
point2(x1, y1, x2, y2, th),
// This point connects this segment to the next
point4(x2, y2, x3, y3, th),
point3(x1, y1, x2, y2, th),
point4(x1, y1, x2, y2, th)
],
paths = [[0,1,2,3,4]]
);
}
else if((x2 > x1 && x2-diffx(x2, y2, x3, y3, th) > x2-diffx(x1, y1, x2, y2, th) || (x2 <= x1 && x2-diffx(x2, y2, x3, y3, th) < x2-diffx(x1, y1, x2, y2, th)))) {
polygon(
points = [
point1(x1, y1, x2, y2, th),
point2(x1, y1, x2, y2, th),
// This point connects this segment to the next
point1(x2, y2, x3, y3, th),
point3(x1, y1, x2, y2, th),
point4(x1, y1, x2, y2, th)
],
paths = [[0,1,2,3,4]]
);
}
else {
polygon(
points = [
point1(x1, y1, x2, y2, th),
point2(x1, y1, x2, y2, th),
point3(x1, y1, x2, y2, th),
point4(x1, y1, x2, y2, th)
],
paths = [[0,1,2,3]]
);
}
}
module 2dgraph(bounds=[-10,10], th=2, steps=10, polar=false, parametric=false) {
step = (bounds[1]-bounds[0])/steps;
union() {
for(i = [bounds[0]:step:bounds[1]-step]) {
if(polar) {
nextPolygon(polarX(i), polarY(i), polarX(i+step), polarY(i+step), polarX(i+2*step), polarY(i+2*step), th);
}
else if(parametric) {
nextPolygon(x(i), y(i), x(i+step), y(i+step), x(i+2*step), y(i+2*step), th);
}
else {
nextPolygon(i, f(i), i+step, f(i+step), i+2*step, f(i+2*step), th);
}
}
}
}
我的包装器代码:
include <2dgraphing.scad>;
include <3dplot.scad>;
function z(x,y) = pow(x,2)+pow(y,2); //function used in 3dplot
function r(theta) = cos(4*theta); //function used in 2dgraph
module Petals () {
difference () {
union () { //everything to add
intersection () {
3dplot([-4,4],[-4,4],[50,50],-2.5);
scale([20, 20, 20]) linear_extrude(height=0.35)
2dgraph([0, 720], 0.1, steps=160, polar=true);
}
}
union () { //everything to subtract
}
}
}
Petals();
当我渲染世界上计算成本最高的花瓣时,世界一切都很好,花花公子。
[这里我会 post 一张图片,但因为这是我的第一张 post 我没有先决条件的 10 个声望点数]
但是,现在我想从花瓣底部减去多余部分。所以我可以使用具有更陡函数和更低起点的 3D 图,并从原始 3D 图中减去它。
所以在同一个程序中我想使用两个不同的函数来实现 3Dplot 模块的两种不同用途。
我尝试修改 3dplot 和我的代码来这样做:
Modified 3dplot:
module 3dplot(x_range=[-10, +10], y_range=[-10,10], grid=[50,50], z_min=-5, dims=[80,80], input_function)
{
dx = ( x_range[1] - x_range[0] ) / grid[0];
dy = ( y_range[1] - y_range[0] ) / grid[1];
// The translation moves the object so that its center is at (x,y)=(0,0)
// and the underside rests on the plane z=0
scale([dims[0]/(max(x_range[1],x_range[0])-min(x_range[0],x_range[1])),
dims[1]/(max(y_range[1],y_range[0])-min(y_range[0],y_range[1])),1])
translate([-(x_range[0]+x_range[1])/2, -(y_range[0]+y_range[1])/2, -z_min])
union()
{
for ( x = [x_range[0] : dx : x_range[1]] )
{
for ( y = [y_range[0] : dy : y_range[1]] )
{
polyhedron(points=[[x,y,z_min], [x+dx,y,z_min], [x,y,input_function(x,y)], [x+dx,y,input_function(x+dx,y)],
[x+dx,y+dy,z_min], [x+dx,y+dy,input_function(x+dx,y+dy)]],
faces=prism_faces_1);
polyhedron(points=[[x,y,z_min], [x,y,input_function(x,y)], [x,y+dy,z_min], [x+dx,y+dy,z_min],
[x,y+dy,input_function(x,y+dy)], [x+dx,y+dy,input_function(x+dx,y+dy)]],
faces=prism_faces_2);
}
}
}
}
修改了我的代码:
include <2dgraphing.scad>;
include <3dplot.scad>;
function z1(x,y) = pow(x,2)+pow(y,2); //function used in 3dplot
function z2(x,y) = pow(pow(x,2)+pow(y,2),1.5)-1; //function to be subtracted out
function r(theta) = cos(4*theta); //function used in 2dgraph
module Petals () {
difference () {
union () { //everything to add
intersection () {
3dplot([-4,4],[-4,4],[50,50],-2.5);
scale([20, 20, 20]) linear_extrude(height=0.35)
2dgraph([0, 720], 0.1, steps=160, polar=true, input_function=z1);
}
}
union () { //everything to subtract
3dplot([-4,4],[-4,4],[50,50],-2.5,input_function=z2);
}
}
}
Petals();
我收到以下错误:
警告:忽略未知函数 'input_function'.
那么我该如何着手将函数作为参数传递?
在此之前我没有用任何函数式语言编写过,但我从 OpenSCAD 用户手册中了解到 "Since version 2015.03, Variables can now be assigned in any scope." 所以我应该能够为每个 input_function 更改值 运行 of 3dplot,就像3dplot中的变量一样。我是不是理解错了?
作为一个可选的附带问题:OpenSCAD 是否有明确的方法来实现我当前的目标,而不会在渲染过程中产生大量计算负载?
我已经花了足够多的时间来尝试解决这个问题,所以我 post 正在回答这个冗长的问题,如果我对一个简单的现有解决方案感到困惑,我深表歉意。我非常感谢任何愿意提供帮助的人。
目前无法将函数作为参数传递。同时生成大量小对象(例如 3dplot 模块中的多面体)会使模型渲染非常慢。为了这
特定用例还有其他选项可以生成模型。
最新 OpenSCAD 版本提供的新列表生成功能允许基于函数生成单个多面体。
查看3d-functions.scad in the demo repository. This plots
函数 f(x, y).
您可以使用我的 function-parsing library 并将相对简单的函数作为文本字符串或作为预编译的优化表示形式传递。该库支持大多数内置的 OpenSCAD 数学和其他函数,但您不能调用其他用户定义的函数,也不能在函数内部进行递归或列表生成。
在过去的几天里,我对使用基于编程语言的软件创建 3D 模型的想法产生了兴趣。我一直在玩的一种语言是 OpenSCAD,事实证明它对创建有趣的形状非常有帮助。
我目前正在尝试使用 OpenSCAD 创建一朵花,我遇到了一个问题,我无法使用我在网上找到的文档或其他资源来规避。
这里是问题的简短形式:
我可以将函数作为参数传递给 OpenSCAD 模块吗?
如果可以,怎么做?如果没有,为什么不呢?我可以做些什么呢?
这让我想到了问题的长篇形式,并根据我的情况做了具体说明:
我正在尝试使用 2D 极坐标函数的线性拉伸创建花瓣,并将其与 3D 函数相交。
为此,我从 http://spolearninglab.com/curriculum/lessonPlans/hacking/resources/software/3d/openscad/openscad_math.html 上找到的两个非常好的模块开始。我并没有声称自己一开始就写了它们。
首先 - Dan Newman 的 3D 绘图仪 /* 3Dplot.scad */
// 3dplot -- the 3d surface generator
// x_range -- 2-tuple [x_min, x_max], the minimum and maximum x values
// y_range -- 2-tuple [y_min, y_max], the minimum and maximum y values
// grid -- 2-tuple [grid_x, grid_y] indicating the number of grid cells along the x and y axes
// z_min -- Minimum expected z-value; used to bound the underside of the surface
// dims -- 2-tuple [x_length, y_length], the physical dimensions in millimeters
//Want to pass in z(x,y) as parameter
module 3dplot(x_range=[-10, +10], y_range=[-10,10], grid=[50,50], z_min=-5, dims=[80,80]){
dx = ( x_range[1] - x_range[0] ) / grid[0];
dy = ( y_range[1] - y_range[0] ) / grid[1];
// The translation moves the object so that its center is at (x,y)=(0,0)
// and the underside rests on the plane z=0
scale([dims[0]/(max(x_range[1],x_range[0])-min(x_range[0],x_range[1])),
dims[1]/(max(y_range[1],y_range[0])-min(y_range[0],y_range[1])),1])
translate([-(x_range[0]+x_range[1])/2, -(y_range[0]+y_range[1])/2, -z_min])
union()
{
for ( x = [x_range[0] : dx : x_range[1]] )
{
for ( y = [y_range[0] : dy : y_range[1]] )
{
polyhedron(points=[[x,y,z_min], [x+dx,y,z_min], [x,y,z(x,y)], [x+dx,y,z(x+dx,y)],
[x+dx,y+dy,z_min], [x+dx,y+dy,z(x+dx,y+dy)]],
faces=prism_faces_1);
polyhedron(points=[[x,y,z_min], [x,y,z(x,y)], [x,y+dy,z_min], [x+dx,y+dy,z_min],
[x,y+dy,z(x,y+dy)], [x+dx,y+dy,z(x+dx,y+dy)]],
faces=prism_faces_2);
}
}
}
}
第二个 - 2D Grapher /* 2dgraphing.scad */
// function to convert degrees to radians
function d2r(theta) = theta*360/(2*pi);
// These functions are here to help get the slope of each segment, and use that to find points for a correctly oriented polygon
function diffx(x1, y1, x2, y2, th) = cos(atan((y2-y1)/(x2-x1)) + 90)*(th/2);
function diffy(x1, y1, x2, y2, th) = sin(atan((y2-y1)/(x2-x1)) + 90)*(th/2);
function point1(x1, y1, x2, y2, th) = [x1-diffx(x1, y1, x2, y2, th), y1-diffy(x1, y1, x2, y2, th)];
function point2(x1, y1, x2, y2, th) = [x2-diffx(x1, y1, x2, y2, th), y2-diffy(x1, y1, x2, y2, th)];
function point3(x1, y1, x2, y2, th) = [x2+diffx(x1, y1, x2, y2, th), y2+diffy(x1, y1, x2, y2, th)];
function point4(x1, y1, x2, y2, th) = [x1+diffx(x1, y1, x2, y2, th), y1+diffy(x1, y1, x2, y2, th)];
function polarX(theta) = cos(theta)*r(theta);
function polarY(theta) = sin(theta)*r(theta);
module nextPolygon(x1, y1, x2, y2, x3, y3, th) {
if((x2 > x1 && x2-diffx(x2, y2, x3, y3, th) < x2-diffx(x1, y1, x2, y2, th) || (x2 <= x1 && x2-diffx(x2, y2, x3, y3, th) > x2-diffx(x1, y1, x2, y2, th)))) {
polygon(
points = [
point1(x1, y1, x2, y2, th),
point2(x1, y1, x2, y2, th),
// This point connects this segment to the next
point4(x2, y2, x3, y3, th),
point3(x1, y1, x2, y2, th),
point4(x1, y1, x2, y2, th)
],
paths = [[0,1,2,3,4]]
);
}
else if((x2 > x1 && x2-diffx(x2, y2, x3, y3, th) > x2-diffx(x1, y1, x2, y2, th) || (x2 <= x1 && x2-diffx(x2, y2, x3, y3, th) < x2-diffx(x1, y1, x2, y2, th)))) {
polygon(
points = [
point1(x1, y1, x2, y2, th),
point2(x1, y1, x2, y2, th),
// This point connects this segment to the next
point1(x2, y2, x3, y3, th),
point3(x1, y1, x2, y2, th),
point4(x1, y1, x2, y2, th)
],
paths = [[0,1,2,3,4]]
);
}
else {
polygon(
points = [
point1(x1, y1, x2, y2, th),
point2(x1, y1, x2, y2, th),
point3(x1, y1, x2, y2, th),
point4(x1, y1, x2, y2, th)
],
paths = [[0,1,2,3]]
);
}
}
module 2dgraph(bounds=[-10,10], th=2, steps=10, polar=false, parametric=false) {
step = (bounds[1]-bounds[0])/steps;
union() {
for(i = [bounds[0]:step:bounds[1]-step]) {
if(polar) {
nextPolygon(polarX(i), polarY(i), polarX(i+step), polarY(i+step), polarX(i+2*step), polarY(i+2*step), th);
}
else if(parametric) {
nextPolygon(x(i), y(i), x(i+step), y(i+step), x(i+2*step), y(i+2*step), th);
}
else {
nextPolygon(i, f(i), i+step, f(i+step), i+2*step, f(i+2*step), th);
}
}
}
}
我的包装器代码:
include <2dgraphing.scad>;
include <3dplot.scad>;
function z(x,y) = pow(x,2)+pow(y,2); //function used in 3dplot
function r(theta) = cos(4*theta); //function used in 2dgraph
module Petals () {
difference () {
union () { //everything to add
intersection () {
3dplot([-4,4],[-4,4],[50,50],-2.5);
scale([20, 20, 20]) linear_extrude(height=0.35)
2dgraph([0, 720], 0.1, steps=160, polar=true);
}
}
union () { //everything to subtract
}
}
}
Petals();
当我渲染世界上计算成本最高的花瓣时,世界一切都很好,花花公子。
[这里我会 post 一张图片,但因为这是我的第一张 post 我没有先决条件的 10 个声望点数]
但是,现在我想从花瓣底部减去多余部分。所以我可以使用具有更陡函数和更低起点的 3D 图,并从原始 3D 图中减去它。
所以在同一个程序中我想使用两个不同的函数来实现 3Dplot 模块的两种不同用途。
我尝试修改 3dplot 和我的代码来这样做:
Modified 3dplot:
module 3dplot(x_range=[-10, +10], y_range=[-10,10], grid=[50,50], z_min=-5, dims=[80,80], input_function)
{
dx = ( x_range[1] - x_range[0] ) / grid[0];
dy = ( y_range[1] - y_range[0] ) / grid[1];
// The translation moves the object so that its center is at (x,y)=(0,0)
// and the underside rests on the plane z=0
scale([dims[0]/(max(x_range[1],x_range[0])-min(x_range[0],x_range[1])),
dims[1]/(max(y_range[1],y_range[0])-min(y_range[0],y_range[1])),1])
translate([-(x_range[0]+x_range[1])/2, -(y_range[0]+y_range[1])/2, -z_min])
union()
{
for ( x = [x_range[0] : dx : x_range[1]] )
{
for ( y = [y_range[0] : dy : y_range[1]] )
{
polyhedron(points=[[x,y,z_min], [x+dx,y,z_min], [x,y,input_function(x,y)], [x+dx,y,input_function(x+dx,y)],
[x+dx,y+dy,z_min], [x+dx,y+dy,input_function(x+dx,y+dy)]],
faces=prism_faces_1);
polyhedron(points=[[x,y,z_min], [x,y,input_function(x,y)], [x,y+dy,z_min], [x+dx,y+dy,z_min],
[x,y+dy,input_function(x,y+dy)], [x+dx,y+dy,input_function(x+dx,y+dy)]],
faces=prism_faces_2);
}
}
}
}
修改了我的代码:
include <2dgraphing.scad>;
include <3dplot.scad>;
function z1(x,y) = pow(x,2)+pow(y,2); //function used in 3dplot
function z2(x,y) = pow(pow(x,2)+pow(y,2),1.5)-1; //function to be subtracted out
function r(theta) = cos(4*theta); //function used in 2dgraph
module Petals () {
difference () {
union () { //everything to add
intersection () {
3dplot([-4,4],[-4,4],[50,50],-2.5);
scale([20, 20, 20]) linear_extrude(height=0.35)
2dgraph([0, 720], 0.1, steps=160, polar=true, input_function=z1);
}
}
union () { //everything to subtract
3dplot([-4,4],[-4,4],[50,50],-2.5,input_function=z2);
}
}
}
Petals();
我收到以下错误: 警告:忽略未知函数 'input_function'.
那么我该如何着手将函数作为参数传递?
在此之前我没有用任何函数式语言编写过,但我从 OpenSCAD 用户手册中了解到 "Since version 2015.03, Variables can now be assigned in any scope." 所以我应该能够为每个 input_function 更改值 运行 of 3dplot,就像3dplot中的变量一样。我是不是理解错了?
作为一个可选的附带问题:OpenSCAD 是否有明确的方法来实现我当前的目标,而不会在渲染过程中产生大量计算负载?
我已经花了足够多的时间来尝试解决这个问题,所以我 post 正在回答这个冗长的问题,如果我对一个简单的现有解决方案感到困惑,我深表歉意。我非常感谢任何愿意提供帮助的人。
目前无法将函数作为参数传递。同时生成大量小对象(例如 3dplot 模块中的多面体)会使模型渲染非常慢。为了这 特定用例还有其他选项可以生成模型。
最新 OpenSCAD 版本提供的新列表生成功能允许基于函数生成单个多面体。
查看3d-functions.scad in the demo repository. This plots 函数 f(x, y).
您可以使用我的 function-parsing library 并将相对简单的函数作为文本字符串或作为预编译的优化表示形式传递。该库支持大多数内置的 OpenSCAD 数学和其他函数,但您不能调用其他用户定义的函数,也不能在函数内部进行递归或列表生成。